lol-2020-orbwalker-opencv-interception-driver-multi-template

Historical 2020 League of Legends orbwalker using OpenCV (multi-template targeting, OCR), DXGI Desktop Duplication capture, and Interception driver for input. External only—no game memory injection. Published as-is; educational only. No build/support. Not affiliated with Riot Games. Explicitly posted as a “personal capability test” and not intended to violate any terms, rights, or policies.

What this is (and why it’s public)

This repository preserves a functional 2020 demo of an external orbwalker prototype. It is shared as an archival study case: computer-vision targeting (template matching + HUD OCR), fast desktop screen capture via DXGI Desktop Duplication, and keyboard/mouse events through the Interception driver.
There is no memory reading/writing of the game. The code is provided as-is for educational/reference purposes only.

No build or usage support. This code is not a product and should not be used to target live services. It exists to document historical techniques and architecture.

Language & naming note

Some variable names, comments, and messages remain in Portuguese on purpose to keep the code as-it-is and historically faithful. Only documentation was standardized to US English.

High-level pipeline

[capture]  -->  [vision]  -->  [core]  -->  [input]
 DXGI/GDI       OpenCV/OCR     timings      Interception
  frames      templates & HUD  (cast/walk)    mouse/keys

• capture: grabs frames from the desktop (DXGI Desktop Duplication, with GDI fallback).
• vision: finds enemies via multi-template matching, thresholds by color, and (optionally) reads stats via OCR.
• core: computes attack windup & cooldown windows (per champion/attack speed) and decides cast vs walk.
• input: tracks the cursor and sends mouse/keyboard events (Space key state, absolute pointer moves).

Directory layout (20 official files)

src/
├─ core/
│  ├─ main.cpp
│  ├─ Globais.cpp
│  ├─ Estruturas.h
│  ├─ GetCastWalkSleep.cpp
│  ├─ GetCastWalkSleep.h
│  ├─ GetScreenSize.cpp
│  └─ GetScreenSize.h
├─ capture/
│  ├─ ScreenCaptureFast.cpp
│  └─ ScreenCaptureFast.h
├─ vision/
│  ├─ GetClosestEnemy.cpp
│  ├─ GetClosestEnemy.h
│  ├─ GetStatsInRealtime.cpp
│  ├─ GetStatsInRealtime.h
│  └─ avir.h
└─ input/
   ├─ GetMousePosInRealTime.cpp
   ├─ GetMousePosInRealTime.h
   ├─ GetSpaceState.cpp
   ├─ GetSpaceState.h
   ├─ interception.c
   └─ interception.h

File-by-file overview

• core/main.cpp — Orchestrates threads/loops and module interaction (demo entry point).
• core/Globais.cpp — Defines global state (cursor positions, timings, champion ID, attack speed, screen size).
• core/Estruturas.h — Basic structs/enums for points, offsets, Interception contexts, etc.
• core/GetCastWalkSleep.[h|cpp] — Computes cast/walk sleep windows from attack speed & champion profiles.
• core/GetScreenSize.[h|cpp] — Detects and exposes supported screen resolutions.
• capture/ScreenCaptureFast.[h|cpp] — DXGI/D3D11 Desktop Duplication for high-FPS capture, with GDI fallback.
• vision/GetClosestEnemy.[h|cpp] — OpenCV color thresholding + multi-template matching to select nearest enemy.
• vision/GetStatsInRealtime.[h|cpp] — HUD parsing; OCR of attack speed/HP (intended for Tesseract/Leptonica).
• vision/avir.h — Header-only image resampler (third-party; used for fast/high-quality resizing).
• input/GetMousePosInRealTime.[h|cpp] — Tracks cursor position (absolute/relative), clamps to screen bounds.
• input/GetSpaceState.[h|cpp] — Polls/holds Space key state via Interception.
• input/interception.[h|c] — Embedded Interception driver user-mode interface (third-party source snapshot).

Historical dependencies (not included)

• Windows 10 (Win32 API).
• DirectX 11 / DXGI 1.2 (Desktop Duplication).
• OpenCV (3.x/4.x) for CV and template matching.
• Tesseract OCR + Leptonica (optional HUD OCR).
• Interception driver for input capture/injection.

This repo does not provide build instructions or binaries. The code is frozen as a historical reference.

Assets & templates (not included)

Some modules expect UI templates such as:

Interface Templates\CHAMPS\ALLY TEMPLATE.png
Interface Templates\CHAMPS\ENEMY TEMPLATE.png
Interface Templates\STATS\*.png

These are not part of the repository and are referenced only to document how the pipeline worked in 2020.

Scope & limitations

• External only (desktop capture + input). No game memory injection.
• No anti-cheat claims are made. This code is not maintained, tested, or intended for live use.
• Timing heuristics, thresholds, template assets, and OCR tuning reflect 2020 experiments.

Ethics & legal

• Educational/archive only. Not affiliated with or endorsed by Riot Games.
• Do not use this to violate ToS, EULAs, anti-cheat policies, or any law.
• Shared explicitly as a “personal capability test” and study case.
• Third-party components remain under their respective licenses (see THIRD_PARTY_NOTICES.md).

Repository housekeeping

• A curated .gitignore is included to keep builds and intermediates out of version control.
• Documentation lives under docs/ (architecture overview, module map, asset references).
• No CI, no issue support, no feature requests.

License

No open-source license is granted for this repository by default (archival release).